Merge remote-tracking branch 'origin/develop' into vpietila/ggemm-profiling

5dee1b64 · Ville Pietilä · 870c3a76 · 355893cd · 5dee1b64 · 5dee1b64
Commit 5dee1b64 authored Dec 09, 2024 by Ville Pietilä
20 changed files
--- a/profiler/include/profiler/profile_batched_gemm_softmax_gemm_impl.hpp
+++ b/profiler/include/profiler/profile_batched_gemm_softmax_gemm_impl.hpp
--- a/profiler/include/profiler/profile_batched_gemm_softmax_gemm_permute_impl.hpp
+++ b/profiler/include/profiler/profile_batched_gemm_softmax_gemm_permute_impl.hpp
--- a/profiler/include/profiler/profile_gemm_impl.hpp
+++ b/profiler/include/profiler/profile_gemm_impl.hpp
--- a/script/process_perf_data.py
+++ b/script/process_perf_data.py
--- a/script/process_perf_data.sh
+++ b/script/process_perf_data.sh
--- a/script/process_qa_data.sh
+++ b/script/process_qa_data.sh
--- a/script/run_full_performance_tests.sh
+++ b/script/run_full_performance_tests.sh
@@ -5,7 +5,7 @@
 # post your new test results to the database and compare them to the baseline
 # please contact Illia.Silin@amd.com for more details
 #
-# run the script as "./run_full_performance_tests.sh <verification> <tag for your test environment> <branch name> < node name>
+# run the script as "./run_full_performance_tests.sh <verification> <tag for your test environment> <branch name> <node name>
 # input arguments: 
 # verification = 0 : do not verify result correctness on CPU
 #              = 1 : verifuy correctness on CPU (may take a long time)

--- a/script/run_gemm_performance_tests.sh
+++ b/script/run_gemm_performance_tests.sh
--- a/script/run_performance_tests.sh
+++ b/script/run_performance_tests.sh
--- a/test/ck_tile/CMakeLists.txt
+++ b/test/ck_tile/CMakeLists.txt
 add_subdirectory(image_to_column)
 add_subdirectory(gemm)
 add_subdirectory(batched_gemm)
+add_subdirectory(grouped_gemm)
--- a/test/ck_tile/gemm/test_gemm_mem_pipeline.cpp
+++ b/test/ck_tile/gemm/test_gemm_mem_pipeline.cpp
--- a/test/ck_tile/gemm/test_gemm_mem_pipeline_ut_cases.inc
+++ b/test/ck_tile/gemm/test_gemm_mem_pipeline_ut_cases.inc
--- a/test/ck_tile/gemm/test_gemm_mem_pipeline_util.hpp
+++ b/test/ck_tile/gemm/test_gemm_mem_pipeline_util.hpp
--- a/test/ck_tile/grouped_gemm/CMakeLists.txt
+++ b/test/ck_tile/grouped_gemm/CMakeLists.txt
--- a/test/ck_tile/grouped_gemm/test_grouped_gemm.cpp
+++ b/test/ck_tile/grouped_gemm/test_grouped_gemm.cpp
--- a/test/ck_tile/grouped_gemm/test_grouped_gemm_ut_cases.inc
+++ b/test/ck_tile/grouped_gemm/test_grouped_gemm_ut_cases.inc
--- a/test/ck_tile/grouped_gemm/test_grouped_gemm_util.hpp
+++ b/test/ck_tile/grouped_gemm/test_grouped_gemm_util.hpp
--- a/test/data_type/CMakeLists.txt
+++ b/test/data_type/CMakeLists.txt
--- a/test/data_type/test_bf8.cpp
+++ b/test/data_type/test_bf8.cpp
--- a/test/data_type/test_bf8_ocp.cpp
+++ b/test/data_type/test_bf8_ocp.cpp